GUEST COLUMN.

For years, organizations have been told that data is their most valuable asset. In the age of artificial intelligence, that familiar phrase needs an important qualification: data may be abundant, but trusted knowledge is becoming increasingly scarce.

The numbers alone are staggering. IDC, a global market intelligence and advisory firm, estimates that the world generated approximately 213.5 zettabytes of data in 2025 and projects that number to reach 527 zettabytes by 2029. A zettabyte is one trillion gigabytes. Put differently, humanity is now creating data at a scale that no individual—and no organization—can hope to manually review, understand, or even meaningfully catalog.

Yet more data does not automatically produce more knowledge.

That distinction may be one of the defining management challenges of the AI era.

From Data to Knowledge

Data consists of raw facts, measurements, records and observations. Information emerges when that data is organized and given context. Knowledge goes a step further: it represents understanding that can be applied to a problem, decision or action.

AI dramatically accelerates our ability to move through those stages. It can summarize thousands of pages, discover patterns across enormous datasets, compare policies, analyze customer interactions, search organizational archives and answer questions in seconds.

But AI is also contributing to the very information explosion it is being asked to solve.

A 2025 analysis by Ahrens, an AI marketing platform, looked at 900,000 newly detected webpages and found that 74.2 percent contained at least some AI-generated content. A more recent study by AI-powered code review and developer productivity platform Graphite, estimated that by early 2026, roughly half of newly published online articles were primarily AI-generated.

This creates what might be called the AI knowledge paradox: AI may be our most powerful tool for managing information overload while simultaneously becoming one of the largest contributors to that overload.

Knowledge Walking Out the Door

There is another knowledge problem that receives considerably less attention.

Employees leave.

They retire. They accept positions elsewhere. They change careers. They are reorganized, downsized, or promoted into new roles. And when they leave, organizations frequently discover that much of what those employees knew was never stored in a database.

The U.S. Bureau of Labor Statistics reported that median employee tenure fell to 3.9 years in 2024, the lowest level since 2002. Among workers ages 25 to 34, median tenure was only 2.7 years. Even among management and professional workers, tenure has been trending downward over the past decade. That matters because some of an organization's most valuable knowledge is tacit, not documented.

Why was a particular policy adopted? Which vendor promised what five years ago? Why does one procedure work differently from what the manual says? Who knows the history behind a contentious decision? What lessons were learned during the last crisis?

When the employee who knows those answers walks out the door, the organization's institutional memory can walk out with them.

AI creates an opportunity to rethink this problem. Organizations can increasingly capture interviews, meeting transcripts, project histories, decision records, policies and lessons learned, then make that knowledge searchable through AI-enabled knowledge systems. The objective should not simply be to archive more documents. It should be to preserve organizational context.

Does the Chief Knowledge Officer Still Matter?

For organizations with a chief knowledge officer (or someone with that responsibility), this raises an intriguing leadership question: Does an organization still need a Chief Knowledge Officer (CKO)? Some might argue that the position has been eclipsed by the Chief Data Officer or Chief Digital Officer. But these roles address different problems.

The Chief Data Officer typically focuses on the quality, governance, accessibility and responsible use of data, while the Chief Digital Officer generally focuses on digital transformation—how technologies can redesign services, operations and customer experiences.

The Chief Knowledge Officer, by contrast, should ensure that an organization can identify, preserve, validate, share, and apply what it knows. With that in mind, far from becoming obsolete, this function may become more important in the AI era. Someone must own the increasingly important space between raw data and organizational judgment. In fact, in the future the CKO, may be less a librarian of institutional information and more a chief curator of organizational intelligence.

That role also includes deciding which knowledge sources AI systems can use, identifying authoritative records, preserving institutional memory, setting verification standards, and helping employees distinguish reliable organizational knowledge from unsupported AI-generated material.

The Biggest Challenge

Separating knowledge from AI junk is critical and may be the most difficult responsibility of all.

AI-generated material can be fluent, polished and persuasive while still being incomplete or wrong. AI detectors themselves are imperfect. The National Institute of Standards and Technology has emphasized that synthetic-content detection, watermarking, authentication and provenance all have roles to play, but each presents technical limitations and implementation challenges.

Organizations therefore cannot solve the problem simply by purchasing an “AI detector.”

They will need something more fundamental: a knowledge chain of trust.

Important organizational knowledge should increase information about its source, ownership, date, authority, and verification status. AI-generated material should be identified where appropriate. Critical facts should link back to authoritative documents or primary sources. Human experts should remain responsible for validating high-consequence information.

In other words, the question should shift from:

“Was this created by AI?”

to:

“Do we know where this came from, and more importantly, can we trust it?”

That distinction is crucial.

Organizations that succeed in making this transition will not necessarily be those that possess the most data or deploy the most AI. They will be the organizations that best preserve what their people know, establish trusted sources, maintain institutional memory, and use AI to connect employees with reliable knowledge when they need it.

In an age of unlimited information, trust may become the most valuable knowledge asset of all.

[Shark’s latest book, “AI for Seniors – A Practical Guide for Living Smarter, Healthier and Safer: in the age of artificial intelligence, can be purchased here.]

The contents of this Guest Column are those of the author, and not necessarily Barrett and Greene, Inc

#StateandLocalGovernmentData #StateandLocalDataGovernance #DistinguishingDataAndKnowledge #StateandLocalGovernmentManagement #StateandLocalPerformance #StateandLocalDataExplosion #StateandLocalKnowledgeManagement #StateandLocalDataKnowledgeChallenge #StateLocalGovernmentGenerativeAIPolicyAndManagement #StateandLocalPerformanceManagement #StateandLocalDataManagement #StateandLocalAI #StateandLocalArtificialIntelligence #ChiefDataOfficer #ChiefDigitalOfficer #ChiefKnowledgeOfficer #StateandLocalDataChallenge #StateandLocalKnowledgeChallenge #CityGovernmentData #CountyGovernmentData #CityKnowledgeManagement #DecreasingEmployeeTenure #CountyKnowledgeManagement #SeparatingKnowledgeFromAIJunk #StateKnowledgeManagement #AlanShark #AlanSharkGuestColumn #AIforSeniors #PublicTechnologyInstitute  #GeorgeMasonUniversity #BandGGuestColumn #BarrettandGreeneInc